We analyze the current annotation of SE constructions, i.e. verbal constructions marked by the clitic se and its cognates across five Romance languages (French, Italian, Portuguese (European and Brazilian), Romanian and Spanish) in several Universal Dependencies treebanks (version 2.17). We discuss the morphologic, syntactic and semantic characteristics of such constructions in each of the languages considered, both from a theoretical perspective and from that of existing annotation. To address inconsistencies in the data and strengthen Universal Dependencies as a scaffold for the automatic conversion of morphosyntactic annotation into semantic representations (Uniform Meaning Representation), we propose a clear distinction between argumental and non-argumental uses of the reflexive clitic, and outline systematic ways to implement this distinction in annotation guidelines. We also examine how some of the reported inconsistencies are handled in the treebanks under study and discuss the extent to which these practices can be extended to other treebanks, within the same or across different languages.

This paper explores the potential of Enhanced and Extended Enhanced Universal Dependencies for narrative analysis in Brazilian Portuguese, focusing on how enhanced dependency relations contribute to clause linkage. The study examines narrative sentences annotated in CoNLL-U format and compares basic and enhanced dependency layers. The results show that enhanced dependencies are particularly useful for recovering implicit subject relations in Portuguese and for enriching clause relations through information from prepositions and conjunctions. These enhancements enable a more explicit representation of participant continuity and event relations in narratives, supporting the transition from syntactic dependency analysis to richer semantic annotation.

Este artigo investiga o modo como os adjetivos representam a experiência visual em descrições, em português brasileiro e em inglês, do conjunto de dados Framed Multi30K. Com base na Gramática Sistêmico-Funcional e na Semântica de Frames, o estudo classifica os adjetivos em Epítetos experienciais, Epítetos interpessoais e Classificadores, além de examinar os frames por eles evocados. Os resultados mostram que os Epítetos experienciais são predominantes, o que indica que os adjetivos são empregados principalmente para representar propriedades visualmente salientes. A anotação semântica também revela que Color é o frame mais frequente. Ao identificar os significados adjetivais mais salientes, o estudo contribui para compreender quais atributos visuais tendem a receber maior destaque por parte dos observadores.

Expressões multipalavras (multiword expressions, MWEs), isto é, sequências de palavras que apresentam algum grau de idiossincrasia, são centrais tanto para os estudos linguísticos quanto para o Processamento de Linguagem Natural (PLN). Este artigo apresenta a criação de um novo corpus de MWEs em postagens em inglês e português brasileiro de comunidades do Reddit dedicadas à saúde mental — um domínio ainda sub-representado nos recursos existentes de MWEs, predominantemente baseados em textos jornalísticos. O corpus foi compilado manualmente, anotado automaticamente com informações morfossintáticas e sintáticas por meio do parser UDPipe, revisado no Arborator Grew e anotado manualmente na plataforma FLAT, seguindo as diretrizes do PARSEME. A análise combina a distribuição de frequência por categoria de MWE e uma investigação contrastiva interlinguística. Os resultados evidenciam especificidades das MWEs nesse domínio e construções que funcionam como recursos convencionalizados para expressar sentimentos, avaliações e estados subjetivos no discurso sobre saúde mental nas redes sociais.

This paper investigates whether sector-specific financial sentiment lexicons improve the analysis of Brazilian stock-market news in Portuguese. We use 19,460 news items about 38 firms in the financial, electric, and manufacturing sectors, published between 2020 and 2025. Starting from OpLexicon, we build sector-specific lexicons through unsupervised expansion based on PMI and conditional word-polarity association. The lexicons are evaluated by lexical heterogeneity and through LSTM price forecasting. Results show sectoral variation and moderate classification performance, but limited predictive gains. Sentiment improves forecasting only in manufacturing, while other sectors do not benefit from its inclusion. The findings indicate that sector-specific lexicons are useful resources, although their predictive value are context dependent.

This paper presents the first Brazilian Portuguese dataset annotated for Uniform Meaning Representation (UMR). It contains 96 sentences from the Brazilian Portuguese portion of the Parallel Universal Dependencies treebank, parallel to existing UMR annotations in English, Czech, and Italian. The sentences were parsed with PortParser, revised in Arborator-Grew, converted from CoNLL-U into preliminary sentence-level UMR graphs, and manually revised in PENMAN format. The results show that CoNLL-U is a useful starting point, but semantic graph construction requires manual interpretation and language-specific lexical resources. The dataset expands multilingual UMR coverage and supports future Portuguese semantic annotation.

This paper introduces a Brazilian Portuguese corpus of fables developed as part of a broader multilingual initiative. It comprises the narratives and their concluding moral statements (epimythia), annotated for morphosyntax and multiword expressions (MWEs) according to the Universal Dependencies (UD) and PARSEME guidelines, respectively. The analysis reveals that 53.59% of the sentences contain at least one MWE, with inherently reflexive verbs (IRVs) constituting the most frequent category. MWEs are also distributed differently: narratives contain a higher proportion of adverbial MWEs, whereas adpositional MWEs are frequent in epimythia. Future work will enrich the corpus through semantic annotation based on Uniform Meaning Representation (UMR).

Uniform Meaning Representation (UMR) is a cross-linguistic semantic representation framework designed to encode sentence meaning in a structured and interpretable way. Building on the foundations of Abstract Meaning Representation (AMR), UMR extends semantic coverage to events, participants, semantic roles, temporal/aspectual information, modality, and discourse links. It is language-agnostic and therefore suitable for multilingual exploration.This tutorial provides a beginner’s introduction to UMR aimed at an audience with no prior experience with AMR, UMR, or meaning representations. The tutorial begins with a simple introduction to the essentials of Universal Dependencies (UD) needed to understand how UMR graphs can be constructed from syntactic information. Using simple Portuguese examples, the tutorial illustrates how basic UD structures guide the creation of UMR graphs. Participants will leave with a foundational understanding of what UMR is; how it relates to syntax and semantic roles; how to create minimal UMR graphs, and how Portuguese UD treebanks can support UMR annotation.

Este artigo apresenta uma avaliação do viés de gênero na tradução automática (TA) do inglês ao português, analisando o desempenho de três tradutores comerciais (Google Translate, Microsoft Translator, Amazon Translate) e três modelos de linguagem de propósito geral (GPT-3.5 Turbo, GPT-4o-mini e Llama-3 8B-Instruct). Utilizando o corpus de teste WinoMT (Stanovsky et al., 2019), a análise quantitativa mediu a acurácia e o viés (ΔG e ΔS) no corpus traduzido. Os resultados mostram que todos os sistemas apresentam viés, com melhor desempenho na tradução de entidades-alvo masculinas (ΔG positivo) e daquelas que corroboram estereótipos ocupacionais (ΔS positivo). A análise qualitativa, fundamentada na Teoria Sistêmico-Funcional, enfocando nas profissões ‘nurse’ e ‘physician’, revela como o viés de gênero constrói significados distintos das sentenças-fontes em relação às entidades-alvo e compromete a coesão referencial. O estudo valida um algoritmo de avaliação adaptado para o português e reitera a persistência do viés como um problema sociotécnico (Savoldi et al., 2025b.). Conclui-se observando a necessidade de avaliações contínuas e de desenvolvimento de métodos de avaliação que considerem diferentes contextos de uso da TA, principalmente em domínios críticos, a fim de ponderar e mitigar danos.

In this paper we analyze the structural and linguistic dynamics of online toxicity in Reddit discussion trees, focusing on how trigger comments escalate conflicts in Brazilian Portuguese. Using a fine-tuned BERTAbaporu model, we show that toxic discussions are deeper, more engaging, and initially semantically cohesive, but degrade over time, while non-toxic interactions emphasize social bonding. Our findings contribute to a better understanding of toxicity escalation and support early detection of discursive conflicts.

Idiomatic expressions present a unique chal-lenge in NLP, as their meanings are often notdirectly inferable from their constituent words.Despite recent advancements in large languagemodels, idiomaticity remains a significant ob-stacle to robust semantic representation. Wepresent datasets and task results for MWE-2026 Shared Task 2: Advancing MultimodalIdiomaticity Representation 2 (AdMIRe 2),which challenges the community to assess andimprove models’ ability to interpret idiomaticexpressions in multimodal contexts across mul-tiple languages. Participants competed in animage ranking task in which, for each item,systems receive a context sentence containinga potentially idiomatic expression (PIE) andfive candidate images. Participating systemsare required to predict the sentence type (i.e.,idiomatic vs. literal) for the given context andrank the images by how well they depict the in-tended meaning in that context. Among the par-ticipating systems the most effective methodsinclude pipelines utilizing closed-source com-mercial models such as Gemini 2.5 and GPT-5, and employing chain-of-thought reasoningstrategies. Methods to mitigate language mod-els’ bias towards literal interpretations and en-sembles to smooth out variance were common.

Potentially idiomatic expressions (PIEs) carry meanings inherently tied to the everyday experience of a given language community. As such, they constitute an interesting challenge for assessing the linguistic (and to some extent cultural) capabilities of NLP systems. In this paper, we present XMPIE, a parallel multilingual and multimodal dataset of potentially idiomatic expressions. The dataset, containing 34 languages and over ten thousand items, allows comparative analyses of idiomatic patterns among language-specific realisations and preferences in order to gather insights about shared cultural aspects. This parallel dataset allows evaluation of language model performance for a given PIE in different languages and whether idiomatic understanding in one language can be transferred to another. Moreover, the dataset supports the study of PIEs across textual and visual modalities, to measure to what extent PIE understanding in one modality transfers or implies in understanding in another modality (text vs. image). The data was created by language experts, with both textual and visual components crafted under multilingual guidelines, and each PIE is accompanied by five images representing a spectrum from idiomatic to literal meanings, including semantically related and random distractors. The result is a high-quality benchmark for evaluating multilingual and multimodal idiomatic language understanding.

We present edition 2.0 of the PARSEME multilingual corpus annotated for multiword expressions (MWEs), resulting from efforts of the PARSEME community towards universality-driven modeling of idiomaticity. With respect to previous editions, we extend the annotation scope to all syntactic MWE categories: verbal, nominal, adjectival, adverbial and functional. We cover 17 languages, of which 7 are new. The annotation process is based on cross-lingually unified guidelines, phrased as decision diagrams over linguistic tests, and a typology of 18 MWE categories. The corpus contains almost 5 million tokens, over 250,000 sentences and 140,000 MWE annotations. The applicability of the corpus is tested in baseline experiments with a prompt-based MWE identification system. Results show that generic large language models do not encode sufficient knowledge to solve the MWE identification task.

This paper details the findings of the 2025 UniDive shared task on multilingual morphosyntactic parsing. It introduces a new representation in which morphology and syntax are modelled jointly to form dependency trees of contentful elements, each characterized by features determined by grammatical words and morphemes. This schema allows bypassing the theoretical debate over the definition of “words” and it encourages development of parsers for typologically diverse languages. The data for the task, spanning 9 languages, was annotated based on existing Universal Dependencies (UD) treebanks that were adapted to the new format. We accompany the data with a new metric, MSLAS, that combines syntactic LAS with F1 over grammatical features. The task received two submissions, which together with three baselines give a detailed view on the ability of multi-task encoder models to cope with the task at hand. The best performing system, UM, achieved 78.7 MSLAS macro-averaged over all languages, improving by 31.4 points over the few-shot prompting baseline.

The increasing complexity of environmental discourse is directly proportional to the growing complexity of environmental debates present today in all communication media. While linguistic and communication studies have been pursued on this discourse, the development of computational linguistic tools and resources dedicated to support its analysis and interpretation is still very incipient. For one, no morphosyntactic resources specific to the environmental domain can be found on major platforms and repositories. This paper introduces TreEn, a multilingual treebank project in progress which compiles texts on environmental discourse produced in different conversational and communication contexts. In particular, it reports on the parallel component of the project and discusses issues faced during sentence-level alignment between original and translated texts, annotation of texts following UD guidelines, and labeling entities drawing on an ontology of environmental-related topics. This novel resource is expected to support environmental discourse analysis by providing morphological and syntactical data to enable cross-language and cross-cultural comparison based on the semantics of the entities annotated in the treebank.

Enhanced Universal Dependencies (EUD) serve as a crucial link between syntax and semantics. Beyond basic syntactic dependencies, EUD provides valuable refined logical connections for downstream tasks such as semantic role labeling, coreference resolution, information extraction, and question answering. The original EUD framework defines six types of relationships, but this paper introduces an extension designed to address subject propagation in pro-drop languages. This “Extended EUD” proposal increases the number of relationships that may be annotated in sentences, improving linguistic representation. Additionally, we report our experiments on a corpus of Portuguese (a pro-drop language), which we make publicly available to the research community.

Contemporary environmental discourse focuses on effectively communicating ecological vulnerability to raise public awareness and encourage positive actions. Hence there is a need for studies to support accurate and adequate discourse production, both by humans and computers. Two main challenges need to be tackled. On the one hand, the language used to communicate about environment issues can be very complex for human and automatic analysis, there being few resources to train and test NLP tools. On the other hand, in the current international scenario, most texts are written in multiple languages or translated from a major to minor language, resulting in different meanings in different languages and cultural contexts. This paper presents a novel parallel corpus comprising the text of World Wide Fund (WWF) 2024 Annual Report in English and its translations into Italian and Brazilian Portuguese, and analyses their linguistic features.

This paper presents a multimodal semantic analysis of accessible Brazilian short films using a frame-based annotation approach. We introduce a subset of the Audition dataset, comprising six short films from the animation and documentary genres. We analysed three communicative modes: original audio, audio description, and visual content. Trained annotators semantically annotated each mode following the FrameNet Brazil multimodal methodology. To compare meaning across modalities, we used cosine similarity over frame-semantic representations. Results show that audio description aligns more closely with video content than original audio, reflecting its role in translating visual meaning into language. Our findings demonstrate the effectiveness of frame semantics in modelling meaning across modalities and provide quantitative evidence of audio description as a bridge between visual and verbal communication. The dataset and annotation strategies are a valuable resource for research on multimodal representation, semantic similarity, and accessible media.

With NLP research being rapidly productionized into real-world applications, it is important to be aware of and think through the consequences of our work. Such ethical considerations are important in both authoring and reviewing (e.g. privacy, consent, fairness, among others). This tutorial will equip participants with basic guidelines for thinking deeply about ethical issues and review common considerations that recur in NLP research. The methodology is interactive and participatory, including discussion of case studies and group work. Participants will gain practical experience on when to flag a paper for ethics review and how to write an ethical consideration section to be shared with the broader community. Most importantly, the participants will be co-creating the tutorial outcomes and extending tutorial materials to share as public outcomes.

This paper presents Framed Multi30K (FM30K), a novel frame-based Brazilian Portuguese multimodal-multilingual dataset which i) extends the Multi30K dataset (Elliot et al., 2016) with 158,915 original Brazilian Portuguese descriptions, and 30,104 Brazilian Portuguese translations from original English descriptions; and ii) adds 2,677,613 frame evocation labels to the 158,915 English descriptions and to the ones created for Brazilian Portuguese; (iii) extends the Flickr30k Entities dataset (Plummer et al., 2015) with 190,608 frames and Frame Elements correlations with the existing phrase-to-region correlations.

This paper presents the Frame2 dataset, a multimodal dataset built from a corpus of a Brazilian travel TV show annotated for FrameNet categories for both the text and image communicative modes. Frame2 comprises 230 minutes of video, which are correlated with 2,915 sentences either transcribing the audio spoken during the episodes or the subtitling segments of the show where the host conducts interviews in English. For this first release of the dataset, a total of 11,796 annotation sets for the sentences and 6,841 for the video are included. Each of the former includes a target lexical unit evoking a frame or one or more frame elements. For each video annotation, a bounding box in the image is correlated with a frame, a frame element and lexical unit evoking a frame in FrameNet.

While Natural Language Processing (NLP) models have gained substantial attention, only in recent years has research opened new paths for tackling Human-Computer Design (HCD) from the perspective of natural language. We focus on developing a human-centered corpus, more specifically, a persona-based corpus in a particular healthcare domain (diabetes mellitus self-care). In order to follow an HCD approach, we created personas to model interpersonal interaction (expert and non-expert users) in that specific domain. We show that an HCD approach benefits language generation from different perspectives, from machines to humans - contributing with new directions for low-resource contexts (languages other than English and sensitive domains) where the need to promote effective communication is essential.

Neural end-to-end surface realizers output more fluent texts than classical architectures. However, they tend to suffer from adequacy problems, in particular hallucinations in numerical referring expression generation. This poses a problem to language generation in sensitive domains, as is the case of robot journalism covering COVID-19 and Amazon deforestation. We propose an approach whereby numerical referring expressions are converted from digits to plain word form descriptions prior to being fed to state-of-the-art Large Language Models. We conduct automatic and human evaluations to report the best strategy to numerical superficial realization. Code and data are publicly available.

We seek to explain the causes of the misclassification of the most challenging documents, namely those that no classifier using state-of-the-art, very semantically-separable contextual embedding representations managed to predict accurately. To do so, we propose a taxonomy of incorrect predictions, which we used to perform qualitative human evaluation. We posed two (research) questions, considering three sentiment datasets in two different domains – movie and product reviews. Evaluators with two different backgrounds evaluated documents by comparing the predominant sentiment assigned by the model to the label in the gold dataset in order to decide on a likely misclassification reason. Based on a high inter-evaluator agreement (81.7%), we observed significant differences between the product and movie review domains, such as the prevalence of ambivalence in product reviews and sarcasm in movie reviews. Our analysis also revealed an unexpectedly high rate of incorrect labeling in the gold dataset (up to 33%) and a significant amount of incorrect prediction by the model due to a series of linguistic phenomena (including amplified words, contrastive markers, comparative sentences, and references to world knowledge). Overall, our taxonomy and methodology allow us to explain between 80%-85% of the errors with high confidence (agreement) – enabling us to point out where future efforts to improve models should be concentrated.

This study describes the development of a Portuguese Community-Question Answering benchmark in the domain of Diabetes Mellitus using a Recognizing Question Entailment (RQE) approach. Given a premise question, RQE aims to retrieve semantically similar, already answered, archived questions. We build a new Portuguese benchmark corpus with 785 pairs between premise questions and archived answered questions marked with relevance judgments by medical experts. Based on the benchmark corpus, we leveraged and evaluated several RQE approaches ranging from traditional information retrieval methods to novel large pre-trained language models and ensemble techniques using learn-to-rank approaches. Our experimental results show that a supervised transformer-based method trained with multiple languages and for multiple tasks (MUSE) outperforms the alternatives. Our results also show that ensembles of methods (stacking) as well as a traditional (light) information retrieval method (BM25) can produce competitive results. Finally, among the tested strategies, those that exploit only the question (not the answer), provide the best effectiveness-efficiency trade-off. Code is publicly available.

This study introduces an enriched version of the E2E dataset, one of the most popular language resources for data-to-text NLG. We extract intermediate representations for popular pipeline tasks such as discourse ordering, text structuring, lexicalization and referring expression generation, enabling researchers to rapidly develop and evaluate their data-to-text pipeline systems. The intermediate representations are extracted by aligning non-linguistic and text representations through a process called delexicalization, which consists in replacing input referring expressions to entities/attributes with placeholders. The enriched dataset is publicly available.

This demo paper introduces DaMata, a robot-journalist covering deforestation in the Brazilian Amazon. The robot-journalist is based on a pipeline architecture of Natural Language Generation, which yields multilingual daily and monthly reports based on the public data provided by DETER, a real-time deforestation satellite monitor developed and maintained by the Brazilian National Institute for Space Research (INPE). DaMata automatically generates reports in Brazilian Portuguese and English and publishes them on the Twitter platform. Corpus and code are publicly available.

This paper introduces the first corpus for Automatic Post-Editing of English and a low-resource language, Brazilian Portuguese. The source English texts were extracted from the WebNLG corpus and automatically translated into Portuguese using a state-of-the-art industrial neural machine translator. Post-edits were then obtained in an experiment with native speakers of Brazilian Portuguese. To assess the quality of the corpus, we performed error analysis and computed complexity indicators measuring how difficult the APE task would be. We report preliminary results of Phrase-Based and Neural Machine Translation Models on this new corpus. Data and code publicly available in our repository.

Data-to-text Natural Language Generation (NLG) is the computational process of generating natural language in the form of text or voice from non-linguistic data. A core micro-planning task within NLG is referring expression generation (REG), which aims to automatically generate noun phrases to refer to entities mentioned as discourse unfolds. A limitation of novel REG models is not being able to generate referring expressions to entities not encountered during the training process. To solve this problem, we propose two extensions to NeuralREG, a state-of-the-art encoder-decoder REG model. The first is a copy mechanism, whereas the second consists of representing the gender and type of the referent as inputs to the model. Drawing on the results of automatic and human evaluation as well as an ablation study using the WebNLG corpus, we contend that our proposal contributes to the generation of more meaningful referring expressions to unseen entities than the original system and related work. Code and all produced data are publicly available.